Wireless Local Area Networks (WLANS) have become ubiquitous. Growth in demand for Wireless Local Area Networks (WLANs) is driving the development of new technology to provide higher throughput. To a greater extent this growth is due to the increased number of users and applications desiring wireless transmission and to a lesser extent to the emergence of new applications needing higher transmission rates along a single connection between two points.
In wireless Local Area Networks (LANs), a wireless channel can be reserved for the transmission of a single frame or of a sequence of frames, known as a TXOP (transmit opportunity), while employing asynchronous distributed random channel access methods, as described in the 2007 802.11 standard, which includes the 11e amendment, where the TXOP was introduced, the IEEE Std 802.11™-2007, (Revision of IEEE Std 802.11-1999), and the Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications.
In such an environment, both the source and destination of the transmission broadcast the reservation duration in order to establish the interference neighborhood. A TXOP is a sequence of frames transmitted between a pair of nodes following a single contention for the channel. A TXOP holder, the node initiating the TXOP, may transmit contention-free after the first transmission to the Responder, which is the node receiving the frames in the TXOP.
To date, there are two basic ways of reserving the channel for a TXOP in wireless LANs. One method of performing reservation is by utilizing Frame-by-frame reservation. According to the 802.11 distributed channel access MAC protocol, RTS/CTS frames are used to notify neighbors of the start of the reservation. Alternatively, the first frame of a TXOP has its duration field set to a time interval long enough to reserve the channel for the transmission of the following frame. The reservation time is extended on a frame-by-frame basis, by updating the length of the reservation with each data frame and the acknowledgement that follows. A consequence of frame-by-frame reservation is that, if the reservation is denied, it does not require cancellation.
Another method of performing reservation in wireless LANS is by utilizing Start-to-finish reservation. If it is not be possible to extend the time of channel reservation on a frame-by-frame basis, the channel must be reserved for the entire sequence of transmissions, start to finish, at the time of the reservation request, and, if the reservation request is denied, or if time remains reserved at the completion of transmission, the reservation must be cancelled.
A start-to-finish reservation applies to any combination of nodes (i.e. mesh points/APs/stations). A node reserves a channel to cover an entire sequence of transmissions, directed to either one or various different destinations, possibly including responses from the destinations. If the reservation is not authorized, or when the transmission sequence is completed, the reserving node releases the remaining reservation time by canceling the reservation.
To avoid collisions, each node keeps a NAV for a traffic channel, which is set according to the received reservation requests and responses. A NAV is defined as a time period a node must refrain from transmitting on a traffic channel. It is maintained by each station and is updated by the Duration field value of received transmissions, which may serve as TXOP reservation requests or responses to reservation requests. A reservation request from the transmissions source is either granted or denied by the destination, and notice is sent to the source. The response contains in the Duration field the remaining reservation duration in order to notify the neighbors of the destination node. Applications for wireless networks include Voice Over Internet Protocol (VoIP), and multimedia (Voice and/or Video), together referred to as VoIP/multimedia. VoIP/multimedia applications require a certain Quality of Service (QoS) in order to maintain sufficient quality of the communication. Latency can be an issue for VoIP/multi-media. Meeting QoS requires short total end-to-end over-the-air delays. The 802.11e amendment to the IEEE 802.11 standard, which is incorporated in the 2007 revision of the standard, provides mechanisms for reducing the over-the-air delays from transmissions in a wireless LAN. These are single hop transmissions. The 802.11e mechanisms may not be adequate for meeting latency requirements in wireless networks involving multiple-hop transmissions. Wireless mesh networks are such networks. A wireless mesh may be an Ad hoc mode mesh (not attached to a wired network) or an infrastructure mode mesh (attached to a wired network). In general, both traffic with source and destination in the mesh and traffic bound from/to a wired network may co-exist on a mesh. The latency/jitter limit for voice traffic traversing the wired network is lower (40 to 50 milliseconds) than that for traffic staying on the wireless mesh (175-200 milliseconds). A mesh will involve multiple-hop flows. The mesh backbone network is a multi-hop network. The multi-hop path delay will be at least a multiple of the single hop delay. Wireless meshes operating on a single channel have novel collision behavior that can impact nearby the latency experienced end to end, over the air. The prevalence of hidden nodes and the interaction of contention-based access with multi-hop flows impose latency increases on both mesh and nearby WLANs beyond what non-mesh experience suggests. Hidden nodes remain hidden after retrial, and their transmissions are dropped. The high correlation of sequentially forwarded frames on a multi-hop flow cause excessive delays to transmissions that have been involved in a collision. For backward compatibility, and for the contention-based access protocol to continue to be used, remedies are needed on the mesh side. For QoS traffic, multi-hop delay must meet the same latency constraints as single-hop delay. We describe remedies to reduce over-the-air latency. The goal is to reduce the delay experienced on the longest multi-hop path by forwarding frames along a multi-hop path fast.
Retransmission of a failed transmission typically involves back off using a wider contention window. By way of express retransmission, a frame is retransmitted by dispensing with back off and transmitting after Short Inter-Frame Space (SIFS) following an ACKTimeout. An express forwarded frame is less likely to collide on retransmission because of its prioritization. Only a first retransmission attempt receives priority treatment in order to prevent two express-forwarded frames from repeatedly colliding.